Looking back at four decades of satellite remote sensing we can safely
say that more research has been conducted into manipulating the spectral
information content of images than any other aspect of digital image analysis
and great progress has been made. Some techniques, such as those associated
with image classification, have undergone continuous improvements throughout,
others like canopy reflectance modelling only emerged in the last decade.
In order to understand why these developments have occurred we require
a better understanding of the nature of information contained in the spectral
domain. The first section of this topic introduces methods to identify
and enhance the spectral component of image data. The second part concentrates
on techniques to extract information.
Before commencing study of this section I would like to draw your attention
to one of the conclusions. It is not possible to declare one approach
better than another. Their performance is strongly data-dependent and,
as with all information products, the quality of the data is ultimately
determined by their fitness-for-purpose. One approach may be more useful
in one application but it is not possible to generalise this statement
to every application. The more experience you can gather the more accomplished
you will become.